14th International Research/Expert Conference

”Trends in the Development of Machinery and Associated Technology”

TMT 2010, Mediterranean Cruise, 11-18 September 2010

OPTIMIZATION OF THE DRYING FACTORS

mr Atif Hodžić
Faculty of technical engineering Bihać ,
University of Bihać
Dr. Irfana Ljubijankića bb, 77000 Bihać
Bosnia & Herzegovina

mr Minka Ćehić mr Ekrem Nezirević

Faculty of technical engineering Bihać , Faculty of technical engineering Bihać ,
University of Bihać University of Bihać
Dr. Irfana Ljubijankića bb, 77000 Bihać Dr. Irfana Ljubijankića bb, 77000 Bihać
Bosnia & Herzegovina Bosnia & Herzegovina

ABSTRACT
In order to improve Kollmann's analytical formula for calculating the duration of the process of
convective drying done experimental research thermodynamic factors of the drying process under real
conditions. Based on the experiment is defined by a mathematical model of the influence of
temperature, humidity and velocity of air in the dryer on the drying of beech planks 32 mm thick, and
the method of linear programming optimal proportions of the same.
Keywords: experiment, model parameters, optimization, temperature, humidity, air velocity.

1. INTRODUCTION
Wood is hygroscopic, anisotropic and orthotropic material that makes it special in operation,
compared with other conventional materials. Drying of wood as an isotropic material difficult by the
fact that the coefficients which determine the thermo-physical properties of wood are not constant in
size, but depend on the direction of extension of grain. Variability drying conditions (temperature,
relative humidity and air velocity) and their dependence on the characteristics, condition, dimensions
and the final exploitation of wood prevents finding an adequate mathematical model of impact
velocity and duration of drying for all cases and all types of wood. Although the theory of heat and
mass transfer in drying wood, as well as physical processes have long known, with the emergence of
numerical analysis and development of computer techniques open the possibility of finding not just a
one-dimensional, but two-dimensional problem of heat transfer.
According to data that can be found in the literature, the general regime of drying connect the relative
humidity and temperature of drying a certain type and thickness of the wood. However, the influence
of relative humidity on the drying rate is not sufficiently explored. The importance of air velocity, as
an influential parameter has not been completely established.
This paper modeled the drying process parameters: temperature, relative humidity and air velocity of
air in the chamber where the dried beech lumber thickness 32 mm, average humidity over 40%, the
final transport humidity. Based on the obtained mathematical model of the same optimization was
performed to determine the optimum drying time for the given conditions.

2. DEFINING THE ISSUES AND CASE STUDIES

Based on theoretical calculation of drying time by Kollmann:

721
 1,25
1 ⎛d⎞ ⎛ 65 ⎞
z = (ln u p − ln uk ) ⎜ ⎟ ⎜ ⎟ (1)
α ⎝ 25 ⎠ ⎝ t ⎠

where:
1,25
⎛ d ⎞
⎜ ⎟ - effect of thickness on the duration of drying,
⎝ 25 ⎠
⎛ 65 ⎞
⎜ ⎟ - effect of air temperature on the duration of drying,
⎝ t ⎠
α - constant drying,
up - initial moisture content of wood,
uk - final moisture content of wood.

The coefficient α is a constant which depends on the conditions of the drying chamber, which are not
adequately tested and determined. It is a constant for constant thickness and bulk density of wood, and
the constant drying conditions. So far, α by a computer equation (2.1), when other sizes are known. This
information as such is still used in the theory and practice, for a specific type of wood and other
conditions of drying. In Kollmann's formula (2.1) there is no humidity, no air flow velocity in the dryer.
Account only the mean temperature, although the temperature increases from start to finish drying. It is
therefore this formula is applicable only for the approximate calculation that meets in practice.
The aim of the research is to model the drying of wood with the aim of optimizing the process
parameters, their impact in the chamber of which depends on the constant α in Kollmann's formula for
calculating the time of drying. The variables of the drying process are temperature, relative humidity
and velocity of air in the dryer. Define the optimal drying process would be improved Kollmann
formula for the duration of the drying process.

3. PLANNING EXPERIMENT
The aim of experimental research to investigate the influence of drying parameters: temperature,
relative humidity and air velocity during artificial drying of beech lumber thickness 32 mm, and the
analysis of experimental data from the standpoint of the potential of achieving high-speed drying, and
shorter duration of the drying process, while at the without compromising the quality of dried
materials and reduce overall production costs. The purpose of the experiment plan is to generate
mathematical models and equations (polynomial of second degree), which describes the process. If
studied parameters in the experiment actually the ones who influence the process and the data obtained
by experiment acceptable accuracy and precision, it is possible to develop a model that describes the
process credible.
Beech lumber is dried in a conventional dryer assembly capacity of 20 m3, which is equipped with an
automatic device for keeping the drying process, supported by computer. Drying is equipped with
functional elements for the maintenance and measurement of study parameters drying temperature (T),
the equilibrium humidity (ur) and currently wood moisture (um), which were duly recorded and stored
in the memory device. To measure the comparative analysis of the actual size of the tested parameters
in the dryer was used to mini data logger with 4 external channels Hobo H8-ext, which with the help
of a temperature sensor TMC-20HD, hygro adapter PCE-HA-702 adapter and anemometer PCE-AM-
402 was measured temperature, relative humidity and velocity of air in the chamber. The device is
equipped with software Boxcar 4.3 for parameterization logger, reading the data, graphing, and the
possibility of comparing several parameters in one graphic. Measurements were performed every hour
during the drying process. Sensors were placed in the chamber, and through flexible cables connected
to the data logger outside the dryer.
The following parameters were investigated drying on five levels:
- Temperature T = 32°, 35°; 40°; 45° and 48 °C
- Relative humidity φ = 42, 45, 50, 55 and 58%
- Air velocity w = 1.1, 1.3, 1.6, 1.9 and 2.1 [m/s]

722
The total number of trials in this experiment was determined with three independently variable size.
Physical and coded values of the parameters of the process of convective drying of wood varied in five
levels. The total number of tests:

N = 2k + nα + n0 = 2k + 2k + n0 = 23 + 2 . 3 + 6 = 20 tests (2)
where:
k = 3 – variable number of variables,
nα= 2k – number of points placed symmetrically around the center of the plan,
n0 = 6 – number of repetitions in the central point plan.

Experimental results show the following diagram:

Figure 1. Experimental results

4. MATHEMATICAL MODELING
In this paper we applied stochastic modeling, with application of active experimental research on a
real object, the convective dryer with intelligent control of drying wood. Pre-plan the experiment was
performed, i.e., varying input parameters of the experiment in the appropriate limits of variation. Used
the regression analysis that determined the functional relationship between the dependent variable in
size y (drying time) and independent variables xi (temperature, relative humidity and air velocity
dryers). This connection can be presented as a function of regression:

y = f ( xi ) (3)
Table 1. Analysis of the adequacy of the mathematical model
Modeling the drying process
of wood made a mathematical
model-tick higher order because of
the previous theoretical analysis
known that the drying time can
represent a linear function. Was used
to model the second order, and
modeling was performed using
rotatable plan that is often applied in
the processes with multiple
variables. After calculating the
regression coefficients and their test
of significance was obtained by a
mathematical model examined the
influence of drying parameters on the duration of the process formalized in the form:

723
 z =z (t,φw)= (–203,37–0,28tφ)–t(0,12t–25,53)+φ(0,08φ–0,53)+w(27,39w–87,68) (4)

5. OPTIMIZATION PROCESS
The optimization procedure is expressed mathematically as requiring minimum or maximum objective
function. Generally, the objective function is a scalar function F with one or more parameters xi,
depending on how many variables are considered during the process of optimizing the model. In
optimization mode processing starts from a mathematical model that describes the process. Set the
appropriate criteria for the optimization and requires a combination of processing elements in the field
of possible solutions to the corresponding objective function given the extreme value (minimum or
maximum).
The paper used the method of nonlinear programming with constraints, direct search, given that the
objective function is not linear but quadratic with iterations between variables. The objective function
(four) set a condition:
Fcilja = y ( x1 , x2 , x3 ) = z3 (t ,ϕ , w)
(5)
and the following limitations:
                                               32,0 ≤ t ≤ 48,0 ; 42,0 ≤ ϕ ≤ 58,0 ; 1,10 ≤ w ≤ 2,10
The programming language C + + program was developed to search the minimum objective function
with the step:
for t: t = t+0,1; for φ: φ = φ+0,1; for w: w = w+0,01.
Limitations are provide in the final points of the interval of modeling parameters: t, φ w, and the
results of several ten thousand combinations of these three parameters were obtained minimum value
of objective function and the parameters t, φ in w in the minimum point of objective function:
Fcilja = z (t ,ϕ , w) = 134,282 = Fmin
respectively:
t = 47,9998 0C ; ϕ = 57,9998 % ; w = 1,60 m / s

6. CONCLUSION
Application of the results obtained in the process of convective drying of wood, and Kollmann
improved formula of the form:
1,25
⎛d⎞
z = zopt ⋅ (ln u p − ln uk ) ⋅ ⎜ ⎟
⎝ 25 ⎠
gives the possibility of more accurate calculation of drying time of beech wood. Justification of
enhanced Kollmann formula in the calculation of the duration of drying of beech compared to
analytical calculations of drying time is as follows:
ƒ An improved model is essentially real as defined in realistic conditions of drying.
ƒ Coefficient of multiple regression R = 0.949 shows a good enough interdependence factors drying
mathematical model, which means that the mathematical model is sufficiently accurate and
reliable is described during the drying process.
ƒ The value of the coefficient of determination R2 = 0.901 means that 90.1% of variability
attributed to action variables (Xi) on the duration of drying wood.

7. REFERENCES
[1] Hodžić A.: „Matematičko modeliranje i optimizacija procesa sušenja u procesu obrade drveta“, doktorska
disertacija, Tehnički fakultet Univrzitata u Bihaću, Bihać, 2010.,
[2] Jurković, M.: „Matematičko modeliranje inženjerskih procesa i sistema“, Mašinski fakultet Bihać, Bihać,
1999.,
[3] Takač A. i sur.: „Matematičko modeliranje“ Prirodno-matematički fakultet Novi Sad, 2006.;
[4] Vukelja D.; Mišković A.: „Inženjerske metode optimizacije“, Građevinska knjiga, Beograd, 1985.

724